KR100467620B1 - Method and apparatus for compressing and/or decompressing binary images - Google Patents

Abstract

The present invention relates to a binary image compression method, and the present invention relates to a method for compressing and / or restoring an image for saving a transmission time and a memory for storing a binary image or transmitting data in a printer, and an apparatus for performing the same. The present invention provides a binary image compression method for generating a record using 2n bits in which a repetition number of 0 and 1 is generated to generate a record. to provide. Accordingly, the present invention improves the method of receiving 4 or 8 bytes from the input stream in the prior art and used it as a record to obtain the repetition number of 0 and 1 of a binary image, and then configure the record in a pair of two repetition numbers to form a bit. By increasing the probability of finding repeating patterns in the dictionary list by allowing more information to be loaded and by reducing the size of the records to less than 4 bytes, the compression rate can be reduced in binary images obtained by mixed glyphs or error diffusion methods. It can be improved.

Description

Binary image compression and / or decompression method and apparatus {Method and apparatus for compressing and / or decompressing binary images}

The present invention relates to a method for compressing and / or restoring a binary image, and more particularly, to a memory for storing or transferring a binary image in a printer, a method of compressing and / or restoring an image for saving transmission time, and an apparatus for performing the same. It is about.

The data compression / restore technology using the conventional Lempel-Ziv-based sliding dictionary is composed as follows.

The video compression part includes all records, dictionaries, window initialization part, mirror rotation of the input stream, input stream of 4 byte or 8 byte unit, and fill in the record. Finds the matching record in the window first, finds the index that matches the current record in the window, prints the index corresponding to the record, finds the dictionary in the window, finds the record in the dictionary, If a matching record is found, the part that outputs the index corresponding to the record, and if the dictionary does not find a record that matches the current record, the part is newly registered in the dictionary list and the marker (xer 0xFE) is added and output as compressed data. When the number of lists reaches 254, the part that replaces the current position in the dictionary with 0 And a part for outputting an end marker (0xFF).

Also, the image restoration part includes all records, dictionaries, window initialization part, one byte input from the input compressed file and checks whether it is an end marker, and ends it if it is an end marker. And a part for outputting bytes or 8 bytes as the restoration data, and a part for finding and outputting a record corresponding to the list number in the dictionary since it is a list number in the dictionary if it is not a marker (Marker 0xFE).

The existing data compression / restore technique is an improvement method of the Lempel-Ziv compression method, and the methods used in the Lempel-Jib compression methods are commonly used. In the Rempeljib compression method, compression occurs when data similar to the input stream is repeated. In the existing compression technology, the conventional Rempel Jib method is extended to set the size of the record to 4 bytes or 8 bytes and to fix the size of the dictionary to shorten the compression and restoration time. In other words, the existing compression method finds a record that is repeated in the dictionary with 4 or 8 byte input streams, and outputs only the list number when a matching record is found. Newly added to and output as compressed data. In addition, in the conventional restoration method, compressed data, which outputs only the list number from the compressed data, searches for and restores the record corresponding to the list number in advance. Otherwise, the compressed data stream following the marker is output as it is. Restore

As described above, in the conventional data compression / restore method, a record is directly input in units of 4 bytes or 8 bytes in an input stream. Using this method to organize records makes it difficult to achieve compression ratios of up to four to one or eight to one. In the case of dithering half-toning image compression, although the correlation between data streams is high, it is difficult to see the gain in the compression ratio due to the limitation of the maximum compression ratio, and the error diffusion method has a relatively low correlation. In the compression of binary images obtained by, it is very difficult to find a record in the dictionary that matches the input record in units of 4 or 8 bytes. Also, unnecessary compression loss occurs due to the size of the marker with more bits allocated than necessary.

Accordingly, an object of the present invention is to solve the above-mentioned disadvantages, to improve the method used as a record by receiving 4 bytes or 8 bytes from the input stream in the prior art to obtain the number of repetitions of 0 and 1 of the binary image By organizing a record into a pair of number of repetitions, it can carry more information per bit, and by reducing the size of the record to less than 4 bytes, it increases the probability of finding a repeating pattern in the dictionary list. The present invention provides a binary image compression method and apparatus for improving a compression ratio in a binary image obtained by a mixed image or an error diffusion method.

Another object of the present invention is to obtain a run length for the list number found in the dictionary list to improve the point that the compression rate in the highly correlated character image has not benefited from the compression rate by limiting the maximum compression rate. In addition, the present invention provides a method and apparatus for compressing a binary image capable of improving a compression ratio even in a video having high correlation between input streams.

Another object of the present invention is to provide a method and apparatus for compressing a binary image which can increase the compression rate by reducing the number of bits unnecessarily allocated to the marker to 1 bit.

Another object of the present invention is to provide a method and apparatus for restoring a binary image compressed by the binary image compression method and apparatus.

1 is a flowchart illustrating image compression according to the present invention.

2 is an exemplary diagram showing a preprocessing method for compressing a binary image obtained by an error diffusion method.

3 is a view for explaining the record generation method of the present invention.

4 is a flowchart illustrating image restoration of the present invention.

5 is a comparison of compression ratios of binary images obtained by a dithering method.

6 is a comparison of compression ratios of binary images obtained by an error diffusion method.

One aspect of the present invention for achieving the above object is a binary image compression method for compressing a binary image by searching for a record matching a current record using a dictionary, and generating a record by obtaining a repetition number of 0 and 1 In addition, the present invention provides a binary image compression method comprising generating a record using 2n bits in which a repetition number n bits for 0 and a repetition number n bits for 1 are combined.

Preferably, the method includes checking whether the image to be compressed is a dithering image or an error diffusion image, and preprocessing the error diffusion image.

More preferably, in the preprocessing step, the portion that changes from white to black and black to white is continuously expressed as '1' or expressed as '0' in a section in which black is maintained to obtain new input data or white In this case, the black and white to black portions are expressed as '0', and in a section where white is maintained or black is '1', new input data is obtained.

Preferably, retrieving records from the dictionary, outputting the current record and adding it to the dictionary when no matching record is found, and counting the number of repetitions of the matching record when the matching record is found, Checking whether the number of repetitions is greater than or equal to two, outputting a reference value and a number of dictionaries when the number of repetitions is two or more, and outputting a reference value of the dictionary when the number of repetitions is less than two.

According to another aspect of the present invention, in a binary image compression apparatus for retrieving a binary image and compressing a binary image by using a dictionary, generating a record by obtaining a repetition number of 0 and 1, The present invention provides a binary image compression apparatus including a compression unit that generates a record using 2n bits in which a repetition number n bits and a repetition number n bits for 1 are combined.

Preferably, the compression unit checks whether an image to be compressed is a dithering image or an error diffusion image, and preprocesses the error diffusion image.

More preferably, the compression unit expresses new input data by expressing a portion that changes from white to black and black to white to '1' continuously in white or '0' in a section maintaining black, or in white The method for obtaining new input data may be performed by expressing black, a portion changed from black to white as '0', and maintaining the white or '1' in a section maintaining black.

Preferably, the compression unit searches for a record in a dictionary, outputs the current record when the record is not found, adds it to the dictionary, calculates the number of repetitions of the matching record when the matching record is found, and the number of repetitions. Is 2 or more, and if the number of repetitions is 2 or more, the reference value and the number of times of the dictionary are output, and if the number of repetitions is less than 2, the dictionary reference value is output.

According to another aspect of the present invention, in a binary image restoration method of restoring a binary image by searching for a record matching a current record using a dictionary, reading the 1 bit of the received binary image and confirming whether the marker is 1; ; Reading the two-byte record to add it to the dictionary when the marker is zero, and splitting the record into four bits to output '0' and '1' by each value; If 1, reads one more byte and checks whether it is 0xFF; If it is not 0xFF, find the record value corresponding to the reference value of the dictionary, divide the record into 4 bits, and output '0' and '1' repeatedly by each value; if 0xFF, read 2 more bytes first Using the value as the reference value of the dictionary and using the second value as the number of iterations, outputting the same method as in the case of not being 0xFF; Checking whether the image is a dithering image or an error diffusion image; Detecting a black and white conversion point in the case of an error diffusion image; Outputting decompressed data in the case of a dither image; And repeating the above steps 401-412 until the file length becomes height * width.

According to another aspect of the present invention, a binary image restoring apparatus for restoring a binary image by searching for a record matching a current record using a dictionary, the method comprising: checking whether a marker is 1 by reading 1 bit of a received binary image; Reading the two-byte record to add it to the dictionary when the marker is zero, and splitting the record into four bits to output '0' and '1' by each value; If 1, reads one more byte and checks whether it is 0xFF; If it is not 0xFF, finding a record value corresponding to a reference value of a dictionary, dividing the record into 4 bits and repeatedly outputting '0' and '1' by each value; In the case of 0xFF, reading two more bytes, using the first value as a reference value of the dictionary, and using the second value as the number of repetitions, and outputting the same method as in the case of not 0xFF; Checking whether the image is a dithering image or an error diffusion image; Detecting a black and white conversion point in the case of an error diffusion image; Outputting decompressed data in the case of a dither image; And a restoration unit for restoring the received binary image by repeating the above steps 401-412 until the file length becomes height * width.

DETAILED DESCRIPTION Hereinafter, embodiments of the structure and operation of the method and apparatus for compressing and restoring a binary image according to one embodiment of the present invention will be described in detail with reference to the accompanying drawings.

In the binary image compression method of the present invention, as shown in FIG. 1, unlike the conventional RempelJib-based data compression method, a step of initializing a dictionary (101), checking whether the input data is empty (102), A step 103 of checking whether it is a dithering image or an error diffusion image, a step of preprocessing 104 in the case of an error diffusion image, a step of generating a record by obtaining a repetition number of 0 and 1, and a step of retrieving from a dictionary 106 Outputting the current record (112) and adding it to the dictionary (113) when a matching record is not found (107), and calculating the number of iterations of the matching record when the matching record is found (107) (108), checking if the repetition number is 2 or more (109), if the repetition number is 2 or more, outputting the reference value and the number of times of the dictionary 110, and outputting the reference value of the dictionary when the repetition number is less than 2; Find Step 111 The.

In the binary image reconstruction method of the present invention, as shown in FIG. 4, a step of initializing a dictionary (400), reading a single bit (401), and checking whether the marker is 1 (402), or the marker is 0 is 2. Read a byte record and add it to the dictionary, split the record into 4 bits, and output '0' and '1' by each value (404) .If 1, read one more byte (403) and check whether it is 0xFF. In step 405, if it is not 0xFF, finding the record value corresponding to the reference value of the dictionary, dividing the record by 4 bits and repeatedly outputting '0' and '1' by each value (407), and 0xFF. In the case of reading two more bytes (406), using the first value as a reference value of the dictionary and the second value as the number of iterations, and outputting in the same manner as in the case of not 0xFF (408), then the dither image Determining whether the image is an error diffusion image (409), Detecting black and white conversion points (410), outputting decompressed data (411) in the case of a dither image, and repeating steps 401-412 described above until the length of the file becomes height * width ( 412.

Hereinafter, the configuration and operation of the above-described binary image compression and decompression method of FIGS. 1 and 4 will be described in detail.

In the present invention, compression occurs when the input data stream has a form in which similar data is repeated based on the Lempel-Ziv method.

1 shows a compression flow chart. First, the dictionary is initialized (101). In the present invention, in order to represent the dictionary list number in 8 bits, it has 253 dictionary list numbers from 2 to 254. 0, 1, and 255 are not included in the dictionary list number for use as markers. By using 0 and 1 as markers, bit loss is minimized. In step 102, it is checked whether the current input stream no longer exists, and terminates when no further input data exists. In the present invention, preprocessing is performed to increase the compression ratio on the binary image obtained by the error diffusion method. In step 103, if the binary image to be compressed is a dithering image or an error diffusion image, it is preprocessed if it is an error diffusion binary image (104). The pretreatment process method is illustrated in FIG. 2. The preprocessing will export '1' at the point where the input stream changes from '0' to '1' and from '1' to '0', or '0' if it keeps exporting the same value as the previous value. The reason for such preprocessing is that in case of error diffusion binary image, a pattern such as '010101 ....' is frequently appeared. When preprocessing using the method shown in FIG. Can generate new high data. In step 105, a record is generated by obtaining the number of repetitions of '0' and '1'. In a mixed image of letters and pictures, '0' or '1' does not repeat continuously long, so if a lot of bits are assigned to the number of repetitions of '0' and '1', the compression rate decreases because of bit waste. In the present invention, bit loss is minimized by limiting the number of repetitions of '0' and '1' to a maximum of 15 times. In the present invention, 8 bits are made by adding 4 bits of '0' and 4 bits of '1'. These two 8-bit data come together to create one record. The size of the record was limited to 2 bytes. In addition, when the size of the record increases in the mixed character picture image, the record having the same pattern becomes difficult to find in the dictionary, so it is limited to 2 bytes. 3 shows a record generation method. In step 106, the record matching the obtained record is found in the dictionary. If no matching record is found, in step 112, the marker '0' and the two-byte record are exported as compressed data and added to the dictionary list (113). When the dictionary list number reaches 254, add the list number 2 again. If a matching record is found in the dictionary in step 107, the value of the dictionary list number is counted in step 108 to obtain a run length for the dictionary list number. If the same dictionary list number is repeated two or more times (109), in the case of two or more times, the marker '1' and '0xFF' and the repeated dictionary list number and the number of repetitions are output as compressed data (110). If it is not more than two times, only the marker '1' and the dictionary list number are exported as compressed data. In the present invention, step 110 compensates for the problems that may occur by limiting the number of repetitions and the size of the records of '0' and '1' at the front. That is, in the case of compressing a character image having a large correlation between binary image data, as the number of repetitions of '0' and '1' increases, the compression ratio becomes large. The present invention provides a maximum compression ratio of 250: 1 for binary images with high correlation. Therefore, in the present invention, the compression ratio is improved in the text image, and the compression ratio is also improved in the mixed text image.

4 is a flowchart of image restoration. First, a dictionary is initialized (400), and a marker of 1 bit is read (401). If the marker is 1 or 0, it is checked (402). If it is 0, no matching record is found in the dictionary, and the record is read from the compressed data and divided into 4 bits. Since the size of the record is 2 bytes, '0' and '1' are repeatedly exported as reconstruction data for each 4-bit value. The record is then added to the dictionary (404). If the marker is 1 in step 402, one byte is read (403) to see if the marker is 0xFF (405). If it is not 0xFF, the dictionary list number is repeated less than two times, so a 2-byte record is obtained from the dictionary list number corresponding to the current 1-byte value. Then, as in step 404, the data is divided into 4 byte units, and the data are repeatedly exported to obtain '0' and '1' (407). In the case of 0xFF in step 405, since the dictionary list number is repeated two or more times, two more bytes are read (406) to obtain the repeated dictionary list value and the number of repetitions. The maximum number of repetitions is 255, which improves the maximum compression ratio for highly correlated binary images. In step 408, the list value may be repeated as many times as in step 407 to generate the restored data. In step 409, it is checked whether the image is a dithering image or an error diffusion binary image. In the case of the error diffusion image, the process proceeds to step 410 and the final reconstruction data is output in the reverse manner to the method shown in FIG. Proceeds to the reverse of the method shown in Figure 2 to output the final reconstruction data. In step 412, the restoration is terminated by checking whether the size of the restoration data is the same as the size of the final restoration data.

In the present invention as described above, the compression ratio is improved not only for the image having a higher correlation between image data such as a character image, but also for a complex image in which characters and pictures obtained by an error diffusion method are mixed. In the compression method of the present invention, it is easy to find a record with a matching pattern by reducing the bit loss and reducing the size of the record to apply to a complex image having a mixture of letters and pictures. The compression rate is improved by obtaining the run length for the dictionary list number for the case of the image with high correlation between the image data. In addition, in the case of error diffusion images, the correlation is increased by preprocessing to increase the compression efficiency, thereby improving the compression ratio. In FIG. 5, the compression ratios of the binary images obtained by the dithering method are compared. In FIG. 6, the compression ratios of the binary images obtained by the erotic diffusion method are compared. The decompression method proposed in the present invention can be used to store and restore images such as patterns and fonts in a printer because the restoration time is several times faster than the compression time, and can also be used for binary image data transmission in a GDI type printer. It is possible.

Although the embodiments of the present invention have been described so far, the present invention is not limited thereto, and various embodiments which can be modified and changed within the true spirit and scope of the principles described and claimed are also included in the scope of protection of the present invention. It should be understood that it belongs.

Claims (10)

In the binary image compression method for compressing a binary image by searching for a record matching a current record using a dictionary, And generating a record by obtaining a repetition number of 0 and 1, and generating a record using 2n bits that combine n bits of 0 and n bits of 1 for repetition. The binary image compression method of claim 1, further comprising: checking whether the image to be compressed is a dithering image or an error diffusion image, and preprocessing the error diffusion image. 3. The method of claim 2, wherein the preprocessing step is to obtain new input data by expressing a portion that changes from white to black and black to white as '1' or '0' in a section maintaining black. Or a step of obtaining new input data by expressing a portion that changes from white to black and black to white as '0' and '1' in a section where white is continuously maintained or black is maintained. 4. The method of any one of claims 1 to 3, further comprising: retrieving a record from a dictionary, outputting the current record and adding it to the dictionary when no matching record is found, and matching when a matching record is found. Calculating the number of repetitions of the record, checking if the number of repetitions is 2 or more, outputting a reference value and the number of dictionaries if the number of repetitions is 2 or more, and outputting a reference value of the dictionary if the number of repetitions is less than 2 Binary image compression method characterized in that it further comprises. In the binary image compression device for compressing a binary image by searching for a record matching the current record using a dictionary, A binary image compression device comprising a compression unit for generating a record by obtaining a repetition number of 0 and 1 and generating a record by combining 2 n bits of repetition number n bits for 0 and n bits of repetition number for 1. 6. The binary image compression apparatus as claimed in claim 5, wherein the compression unit checks whether the image to be compressed is a dithering image or an error diffusion image, and preprocesses the error diffusion image. The method of claim 6, wherein the compression unit expresses a new portion of white to black, a portion changed from black to white to '1' to maintain white or to '0' in a period of maintaining black, or Binary image compression device, characterized in that the step of obtaining new input data by expressing the portion that changes from white to black and black to white as '0' and '1' in the period of maintaining white or black is maintained. . 8. The method according to any one of claims 5 to 7, wherein the compression unit searches for a record in the dictionary, outputs the current record and adds it to the dictionary when no matching record is found, and matches when a matching record is found. It calculates the number of repetitions of the records to check, checks if the number of repetitions is 2 or more, outputs the reference value and the number of the dictionary when the repetition number is 2 or more, and outputs the reference value when the repetition number is less than 2 Binary image compression device. In the binary image restoration method of restoring a binary image by searching for a record matching the current record using a dictionary, Checking whether the marker is 1 by reading 1 bit of the received binary image; Reading the two-byte record to add it to the dictionary when the marker is zero, and splitting the record into four bits to output '0' and '1' by each value; If 1, reads one more byte and checks whether it is 0xFF; If it is not 0xFF, finding a record value corresponding to a reference value of a dictionary, dividing the record into 4 bits and repeatedly outputting '0' and '1' by each value; In the case of 0xFF, reading two more bytes, using the first value as a reference value of the dictionary, and using the second value as the number of repetitions, and outputting the same method as in the case of not 0xFF; Checking whether the image is a dithering image or an error diffusion image; Detecting a black and white conversion point in the case of an error diffusion image; Outputting decompressed data in the case of a dither image; And And repeating the above steps 401-412 until the file length becomes height * width. In the binary image restoration apparatus for restoring a binary image by searching for a record matching the current record using a dictionary, Checking whether the marker is 1 by reading 1 bit of the received binary image; Reading the two-byte record to add it to the dictionary when the marker is zero, and splitting the record into four bits to output '0' and '1' by each value; If 1, reads one more byte and checks whether it is 0xFF; If it is not 0xFF, finding a record value corresponding to a reference value of a dictionary, dividing the record into 4 bits and repeatedly outputting '0' and '1' by each value; In the case of 0xFF, reading two more bytes, using the first value as a reference value of the dictionary, and using the second value as the number of repetitions, and outputting the same method as in the case of not 0xFF; Checking whether the image is a dithering image or an error diffusion image; Detecting a black and white conversion point in the case of an error diffusion image; Outputting decompressed data in the case of a dither image; And And a restoring unit for restoring the received binary image by repeating the above steps 401-412 until the file length becomes height * width.